Effects of Pitavastatin on Insulin Sensitivity and Liver Fat: A Randomized Clinical Trial

Abstract

Context

3-Hydroxy-3-methyl-glutaryl-coenzyme A reductase inhibitors (statins) are widely prescribed. Statins may have important metabolic effects on insulin sensitivity and liver fat, but limited studies have assessed these effects by using euglycemic hyperinsulinemic clamp, stable isotopes, and ¹H magnetic resonance spectroscopy (MRS) for liver fat quantification.

Objective

To study the effects of pitavastatin on hepatic fat and insulin sensitivity.

Design

Six-month, double-blind, randomized, placebo-controlled trial.

Setting

Academic clinical research center in Boston, Massachusetts.

Participants

Overweight, insulin-resistant men aged 40 to 65 years who had not received statin therapy for ≥1 year.

Interventions

Pitavastatin 4 mg or placebo daily.

Outcome

The primary endpoints were changes in insulin sensitivity measured by euglycemic hyperinsulinemic clamp and liver fat measured by ¹H MRS.

Results

Pitavastatin showed no effect on endogenous glucose production (ΔRa glucose 0.07 ± 0.07 vs 0.04 ± 0.07 mg/kg/min, pitavastatin vs placebo, P = 0.76) or insulin-stimulated glucose uptake during “low dose” (ΔM 0.1 ± 0.1 vs −0.3 ± 0.2 mg/kg/min, P = 0.11) and “high dose” (ΔM −0.5 ± 0.3 vs −0.7 ± 0.4 mg/kg/min, P = 0.70) euglycemic hyperinsulinemic clamps. There was also no effect of pitavastatin on fasting glucose, HbA1c, and 2-hour glucose after 75-g glucose challenge. There was also no change in liver fat fraction (−1 ± 1 vs −0 ± 1%, P = 0.56).

Conclusion

Compared with placebo, pitavastatin did not affect hepatic or whole-body insulin sensitivity, and it did not reduce liver fat.

In a double-blind randomized controlled trial, 6 months of pitavastatin treatment did not change hepatic or whole-body insulin sensitivity or reduce liver fat in abdominally obese men with insulin resistance.

Statins, or 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase inhibitors, have pleiotropic effects beyond inhibition of the cholesterol biosynthetic pathway (1). Statin therapy has the clear benefit of reducing cardiovascular mortality independent of lipid-lowering effects (2, 3). Statins are known to play a role in hepatic response to insulin, and some statins increase the risk of new-onset diabetes (4–6). This diabetogenic effect appears to be dose dependent, with a higher risk of diabetes with more intensive dose therapy (7). This effect does not seem to be sustained among all medications in this class, and it appears that some statins such as pravastatin and lovastatin may be neutral to glucose metabolism (8, 9). A key question on the metabolic effects of statins relates to the effects of statins on hepatic fat. Animal studies and small human studies have suggested that statin therapy may decrease hepatic steatosis and steatohepatitis (10–14). Similarly, limited animal and human data show potential effects of statins on liver enzymes (15, 16) and inflammatory indices (17, 18), as well as hepatic fat scores (15, 16). Taken together, these data suggest potentially important effects of statins on metabolic variables. However, previous randomized, placebo-controlled trials in humans have not simultaneously assessed effects of statins on key indices of muscle and hepatic insulin sensitivity by using state-of-the-art euglycemic clamp and stable isotope methods or assessed liver fat by using established magnetic resonance spectroscopy (MRS) imaging techniques.

Pitavastatin is a moderate-intensity, mildly lipophilic statin that short-term studies suggest may have neutral or beneficial effects on glucose metabolism (19, 20). Based on this evidence, as well as preliminary data suggesting that statins may reduce liver fat, we conducted the current study with two aims: to carefully characterize the effects of pitavastatin on hepatic and whole-body insulin sensitivity among insulin-resistant, abdominally obese men, with the hypothesis that pitavastatin would be neutral to insulin sensitivity even among insulin-resistant patients, and to determine whether pitavastatin treatment over 6 months reduces liver fat in this population, with the hypothesis that pitavastatin would decrease liver fat compared with placebo. In this randomized, placebo-controlled trial, we demonstrated that, after 6 months of treatment with pitavastatin, there was no effect on hepatic or whole-body insulin resistance or on liver fat content in the population studied.

Methods

Participants

Men between the ages of 40 and 65 years with abdominal obesity and indications of insulin resistance were recruited from the greater Boston area via the clinical research study volunteer program and advertisements in print and electronic media. Ninety-five potential participants were screened at Massachusetts General Hospital (MGH) between March 2015 and October 2017 (Fig. 1). Written informed consent was obtained from each participant before the screening procedures begin. The first participant was enrolled on 9 March 2015, and the final study visit was completed on 27 April 2018. The study was approved by the MGH Institutional Review Board.

Figure 1.

CONSORT flow diagram.

Subjects were selected to be eligible if they demonstrated both of the following: abdominal obesity, as demonstrated by body mass index (BMI) ≥27 kg/m² and waist circumference ≥102 cm; and prediabetes, defined by either a fasting blood glucose ≥100 mg/dL and <126 mg/dL or a 2-hour glucose ≥140 mg/dL and <200 mg/dL after standard 75-g oral glucose tolerance test (OGTT), or insulin resistance defined by a homeostasis model assessment of insulin resistance (HOMA-IR) (21) ≥2.0. Subjects were also required to have either 10-year atherosclerotic cardiovascular disease (ASCVD) risk ≥5% by American Heart Association guidelines (22) or low-density lipoprotein cholesterol (LDL-C) ≥100 mg/dL, such that treatment with a statin would be clinically reasonable. Subjects with known ASCVD, LDL-C ≥190 mg/dL, or triglyceride ≥500 mg/dL were excluded, as were subjects who were actively being considered for statin therapy by their physician, because randomization to placebo would not have been acceptable in these cases. Other exclusion criteria were as follows: use of statin in the past year, diagnosis of diabetes or use of antidiabetic medications, use of lipid-modifying therapy, renal disease, alanine aminotransferase (ALT) or aspartate aminotransferase (AST) at least three times the upper limit of normal, creatinine greater than the upper limit of normal, or hemoglobin <10 g/dL.

Study procedures

Screening procedures occurred after a 12-hour overnight fast and included anthropometric measurements, fasting lipid panel, fasting insulin levels, ALT, AST, and a standard 75-g OGTT.

Eligible participants were randomly assigned 1:1 to receive pitavastatin calcium 4 mg vs matching placebo by mouth daily for 6 months. The pitavastatin 4 mg and identical-appearing placebo were supplied by Kowa Pharmaceuticals America, Inc. (Montgomery, AL) and packaged by the MGH Research Pharmacy. Randomization was performed by a biostatistician via block randomization with blocks of varying sizes, and the randomization key was shared only with the pharmacy. All study staff and participants were blinded to treatment assignment until study completion.

Participants underwent a baseline visit that included ¹H MRS and stable isotope and clamp assessment, as described here. Participants returned for safety visits at 1 and 3 months and then returned for a final assessment at 6 months that was identical to the baseline assessment and also included repeat OGTT (performed initially at the screening visit). At baseline and at 3 and 6 months, all participants received standardized nutrition and activity counseling by research bionutritionists, who were also blinded to randomization (23, 24).

Euglycemic hyperinsulinemic clamp with stable isotopes at baseline and 6 months

After a 12-hour overnight fast, patients underwent euglycemic hyperinsulinemic clamp procedure under low-dose insulin (20 mU/m²/min) and high-dose insulin (80 mU/m²/min) conditions for 120 minutes each. Patients received an infusion of 20% dextrose at a variable infusion rate to maintain a blood glucose level of 90 mg/dL. Arterialized venous blood was sampled every 5 minutes to measure blood glucose concentration with a point-of-care glucose analyzer (Glucose 201; Hemocue, Cypress, CA). Blood samples were collected for insulin levels at times 0, 80, 100, 120, 200, 220, 240 minutes after the start of the clamp procedure. Hepatic and peripheral insulin-stimulated glucose disposals (M) were calculated during the last 20 minutes of low-dose and high-dose clamp, respectively, with the DeFronzo method (25), as the primary indices of insulin sensitivity. Similarly, M normalized to serum insulin concentration (M/I) was calculated.

A stable isotope infusion of 6,6-²H-glucose (Cambridge Isotope Laboratories, Tewksbury, MA) was infused (3.0 mg/kg priming dose and 0.03 mg/kg/min infusion) starting 2 hours before the start of the clamp. Samples were collected for isotopic analysis (Metabolic Solutions, Nashua, NH) for assessment of fasting endogenous glucose production (Ra glucose) with the modified Steele equation (26).

MRS at baseline and 6 months

¹H MRS was performed after 12-hour overnight fast for quantification of hepatic fat (27). Visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) area were also measured cross-sectionally at L4 with MRI. All imaging was completed on the same scanner (Siemens Trio-3 Tesla; Siemens Medical, Erlangen, Germany), and all scans were reviewed by the same radiologist, who remained blinded to treatment assignment.

Other assessments

Patients underwent laboratory assessment for fasting glucose, insulin, lipid panel, direct LDL-C, HbA1c, ALT, AST, creatine phosphokinase (CPK), and coenzyme Q10 (CoQ10). Glucose, lipid panel, transaminases, and HbA1c were measured by standard clinical assays (Labcorp, Burlington, NC). CoQ10 was measured by liquid chromatography (Labcorp). Serum insulin during the clamp was measured by chemiluminescent assay (Beckman Coulter, Fullerton, CA). Physical activity was evaluated with the Modifiable Activity Questionnaire (28). Food and alcohol intake were quantified with a 3-day food record (Nutrition Data System, Minneapolis, MN). Dual-energy x-ray absorptiometry (Hologic, Discovery A, Marlborough, MA) was performed for total body and regional fat mass (29).

Statistical analysis

The study was powered for detection of clinically significant differences in the coprimary endpoints of changes in insulin-stimulated glucose uptake and hepatic fat fraction. More specifically, the protocol was designed with the plan to enroll 50 participants, with an expected 10% discontinuation rate, which would provide 80% power to detect a change of ≥1.5 mg/kg/min in whole-body insulin sensitivity by using preliminary data, or a change of ≥0.86 SD in hepatic fat fraction at a 2-sided α = 0.05.

The analysis was based on intention to treat, with all available data. Liver fat MRS data from one participant in the placebo group were excluded before unblinding because of inadequate spectral quality. One participant in the placebo group had 6-month insulin values that were outliers by >5 SD; his 6-month insulin values were excluded from analysis. One subject in the pitavastatin group developed a metastatic malignancy in his abdomen before the final visit. Per intention to treat, this subject was retained in the primary analysis, and a sensitivity analysis was also performed excluding him.

Baseline comparisons were expressed as mean ± SEM and compared between groups with the Student two-sample t test. Categorical data comparisons were assessed via the likelihood ratio χ² test. For comparison of outcomes between the pitavastatin and placebo groups, Student two-sample t test was performed assessing the difference in mean change from baseline to 6 months. Data analysis was performed with JMP Pro, version 13.0.0 (SAS Institute Inc., Cary, NC).

Results

Participant flow

The CONSORT flow diagram is shown in Fig. 1. A total of 95 participants were assessed for eligibility; 50 participants were randomly assigned to receive pitavastatin (n = 25) or placebo (n = 25). One participant in each group was lost to follow-up. One participant was discontinued at the discretion of the study physicians because of concerns about myopathy; however, subsequent testing revealed normal CPK.

Study subjects

Subjects were abdominally obese (mean ± SEM, BMI 36.1 ± 0.8 kg/m², waist circumference 121 ± 2 cm), with a high VAT area (271 ± 13 cm²). They were insulin resistant (HOMA-IR 3.0 ± 0.2) and had a high liver fat content (16% ± 2%). Mean 10-year ASCVD risk score was 7.1% ± 0.6%, and mean LDL-C was 133 ± 3 mg/dL. No patient reported current alcohol abuse, and mean reported alcohol intake was 9 ± 2 g/day, which is less than one drink per day. Groups were well matched in terms of demographics, smoking status, antihypertensive status, and body composition (Tables 1 and 2). The participants in the treatment group had slightly lower total cholesterol levels at baseline, a difference that was not clinically significant (182 ± 4 vs 196 ± 6; pitavastatin vs placebo, P = 0.05). The groups were also comparable with respect to all metabolic and liver parameters, with no significant differences between the groups (Tables 1 and 2).

Table 1.

Baseline Characteristics of the Cohort by Treatment Assignment

	Pitavastatin (n = 25)	Placebo (n = 25)
Demographics
Age, y	52.8 ± 1.3	52.9 ± 1.4
Race, n (%)
White	24 (96)	21 (84)
Nonwhite	1 (4)	3 (12)
Not reported	0	1 (4)
Current smoker, n (%)	4 (16)	4 (16)
Antihypertensive use, n (%)	9 (36)	5 (20)
Lipid profile
Total cholesterol, mg/dL	182 ± 4	196 ± 6a
LDL-C, mg/dL	130 ± 4	137 ± 5
HDL-C, mg/dL	41 ± 1	42 ± 2
Triglycerides, mg/dL	132 ± 12	152 ± 14
Metabolic parameters
SBP, mm Hg	135 ± 3	136 ± 3
DBP, mm Hg	83 ± 2	80 ± 2
Fasting glucose, mg/dLb	91 ± 2	91 ± 1
2-h glucose OGTT, mg/dLb	139 ± 6	131 ± 7
HOMA-IRb	2.7 ± 0.3	3.2 ± 0.3
Creatinine, mg/dL	0.81 ± 0.03	0.88 ± 0.02

Continuous data reported as mean ± SEM.

Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure.

^a P = 0.05 (all other baseline comparisons between treatment groups P > 0.05).

^bFasting glucose, 2-h glucose, and HOMA-IR as assessed at screening visit as part of eligibility criteria.

Table 2.

Baseline Values and Outcomes After 6 Mo of Treatment

	Baseline		6 Mo		Change Over 6 Mo
	Pitavastatin	Placebo	Pitavastatin	Placebo	Pitavastatin (n = 24)	Placebo (n = 23)	Effect Size (95% CI)	P
Body composition
Weight, kg	108.6 ± 3.6	116.4 ± 4.8	107.6 ± 3.6	117.4 ± 4.6	−1.0 ± 0.8	1.0 ± 0.8	−2.0 (−4.4, 0.3)	0.09
BMI, kg/m2	35.6 ± 1.0	36.1 ± 1.4	35.4 ± 1.0	36.4 ± 1.3	−0.3 ± 0.3	0.3 ± 0.3	−0.6 (−1.4, 0.1)	0.10
Waist circumference, cm	118 ± 2	122 ± 3	117 ± 2	123 ± 3	−0 ± 1	1 ± 1	−1 (−4, 2)	0.40
VAT area, cm2	272 ± 16	271 ± 24	267 ± 16	287 ± 25	−6 ± 15	16 ± 13	−22 (−62, 19)	0.29
SAT area, cm2	393 ± 25	425 ± 36	405 ± 35	424 ± 33	11 ± 14	−1 ± 12	13 (−25, 51)	0.50
VAT/SAT ratio	0.75 ± 0.08	0.70 ± 0.08	0.78 ± 0.10	0.73 ± 0.08	0.02 ± 0.08	0.04 ± 0.05	−0.01 (−0.20, 0.17)	0.88
Body fat, %	35.8 ± 1.0	36.7 ± 1.2	35.8 ± 0.9	37.2 ± 1.1	0.1 ± 0.4	0.4 ± 0.3	−0.4 (−1.4, 0.6)	0.47
Lipid profile
Total cholesterol, mg/dL	183 ± 4	199 ± 6	137 ± 4	197 ± 6	−46 ± 4	−2 ± 5	−44 (−57, −32)	<0.0001
LDL-C, mg/dL	116 ± 4	125 ± 5	75 ± 3	126 ± 4	−49 ± 3	−1 ± 4	−48 (−58, −37)	<0.0001
HDL-C, mg/dL	41 ± 2	43 ± 2	39 ± 1	43 ± 2	−2 ± 1	−0 ± 1	−1 (−4, 2)	0.38
Triglycerides, mg/dL	133 ± 12	153 ± 15	123 ± 19	144 ± 12	−10 ± 10	−9 ± 11	−1 (−31, 29)	0.95
Metabolic parameters
HbA1c, %	5.7 ± 0.1	5.7 ± 0.1	5.7 ± 0.1	5.7 ± 0.1	0.0 ± 0.1	−0.0 ± 0.0	0.1 (−0.1, 0.2)	0.24
Fasting glucose, mg/dL	98 ± 2	97 ± 2	96 ± 3	99 ± 2	−2 ± 2	2 ± 2	−4 (−10, 1)	0.13
2-hr OGTT glucose, mg/dLa	139 ± 6	132 ± 8	141 ± 9	138 ± 8	3 ± 6	7 ± 6	−3 (−20, 14)	0.71
HOMA-IR	2.1 ± 0.4	1.6 ± 0.2	1.6 ± 0.3	2.0 ± 0.3	−0.5 ± 0.3	0.3 ± 0.2	−0.8 (−1.6, -0.1)	0.03
Fasting insulin, μU/mL	8.5 ± 1.5	6.7 ± 0.8	7.0 ± 1.0	7.9 ± 1.2	−1.9 ± 1.1	1.1 ± 1.0	−3.1 (−6.1, -0.0)	0.05
Low-dose clamp M, mg/kg/min	1.4 ± 0.2	1.9 ± 0.2	1.5 ± 0.2	1.6 ± 0.2	0.1 ± 0.1	−0.3 ± 0.2	0.4 (−0.1, 0.8)	0.11
High-dose clamp M, mg/kg/min	6.4 ± 0.4	6.6 ± 0.5	5.9 ± 0.4	5.9 ± 0.3	−0.5 ± 0.3	−0.7 ± 0.4	0.2 (−0.9, 1.3)	0.70
Ra glucose, mg/kg/minb	1.57 ± 0.06	1.55 ± 0.05	1.66 ± 0.08	1.58 ± 0.07	0.07 ± 0.07	0.04 ± 0.07	0.03 (−0.17, 0.23)	0.76
CPK, U/L	135 ± 18	197 ± 37	137 ± 14	194 ± 34	2 ± 13	−3 ± 24	5 (−50, 59)	0.87
CoQ10, μg/mL	1.0 ± 0.1	1.1 ± 0.1	0.6 ± 0.0	1.0 ± 0.1	−0.4 ± 0.1	−0.1 ± 0.1	−0.4 (−0.7, −0.1)	0.01
Liver parameters
ALT, IU/L	29 ± 3	25 ± 3	37 ± 8	27 ± 3	8 ± 8	3 ± 2	6 (−11, 22)	0.49
AST, IU/L	24 ± 2	24 ± 2	31 ± 5	26 ± 3	7 ± 5	2 ± 2	5 (−5, 15)	0.35
Liver fat content, %c	17 ± 2	14 ± 3	17 ± 2	14 ± 2	−1 ± 1	−0 ± 1	−1 (−5, 3)	0.56

Data reported as mean ± SEM. P for Student two sample t test (two-tailed) comparing changes between groups. Data shown are for patients who had both baseline and follow-up data available.

^aOGTT performed at screen visit and not repeated at baseline. Screen values shown for baseline. Change in 2-h glucose between screen visit and 6 mo.

^b10 participants (5 pitavastatin, 5 placebo) did not have data for stable isotope procedure because of inability to obtain venous access (N = 1) or temporary unavailability of glucose isotope (N = 9).

^c7 participants (2 pitavastatin, 5 placebo) were unable to complete MRI at baseline because of inability to fit in the scanner or unanticipated claustrophobia.

Effects of pitavastatin on glucose metabolism

Treatment with pitavastatin had no effect on fasting hepatic gluconeogenesis (Ra glucose) or insulin sensitivity during low-dose or high-dose euglycemic hyperinsulinemic clamp studies (Table 2, Fig. 2). Euglycemic hyperinsulinemic clamp procedures had adequate low-dose (minutes 220 to 240) and high-dose (minutes 340 to 360) steady states achieved at blood glucose levels of 90 mg/dL at baseline and final visits (Fig. 3A–3D). Insulin-stimulated glucose uptake (M) is shown in Table 2. Normalizing glucose utilization to lean body mass instead of total body weight yielded highly similar results. When we adjusted for serum insulin concentrations, results similarly showed no effect of pitavastatin (low-dose clamp ΔM/I −0.8 ± 0.7 vs −1.0 ± 0.9 mg/kg/min per μU/mL × 100, P = 0.87; high dose clamp ΔM/I 0.3 ± 0.5 vs −0.9 ± 0.5 mg/kg/min per μU/mL × 100, P = 0.10, pitavastatin vs placebo, respectively) (Fig. 2). There were also no effects of pitavastatin on HbA1c, fasting glucose, or 2-hour glucose after OGTT. There was a modest, statistically significant reduction in HOMA-IR in the pitavastatin group compared with the placebo group and a trend toward improvement in fasting insulin (Table 2).

Figure 2.

Box and whisker plots showing change from baseline in (A) LDL-C, (B) low-dose clamp M, (C) high-dose clamp M, (D) hepatic fat fraction, (E) low-dose clamp M adjusted for insulin, and (F) high-dose clamp M adjusted for insulin. Data are presented as median (line), 25th and 75th percentiles (box), extremes (whiskers), and outliers (dots). M is insulin-stimulated glucose uptake.

Figure 3.

(A) Baseline and (B) final visit plasma glucose (diamonds) and insulin (triangles, dotted line) concentrations, and (C) baseline and (D) final visit glucose infusion rates during low-dose (min 0 to 120) and high-dose (min 120 to 240) hyperinsulinemic euglycemic clamp. Data are represented as means and SEM (bars). GIR, glucose infusion rate.

Effects of pitavastatin on liver parameters

There were no significant changes in liver fat content (−1% ± 1% vs −0% ± 1%, pitavastatin vs placebo, P = 0.56, Fig. 2) between the pitavastatin and placebo group or in AST or ALT (Table 2).

Effects of pitavastatin on lipids, body composition, and other parameters

As expected, pitavastatin decreased LDL-C substantially compared with placebo (−49 ± 3 vs −1 ± 4 mg/dL, P < 0.0001) and also reduced total cholesterol (Table 2). High-density lipoprotein cholesterol (HDL-C) and triglyceride did not significantly change. Pitavastatin did not change body weight or any measures of body fat distribution (Table 2). Pitavastatin did not change CPK levels. Compared with placebo, there was a reduction in CoQ10 levels with pitavastatin (Table 2). There were no differences in changes in caloric or macronutrient intake over the study between the pitavastatin and placebo groups (Δcaloric intake 6 ± 184 vs −116 ± 187 kcal/day, P = 0.65; Δfat intake 0 ± 10 vs 7 ± 9 g/day, P = 0.63; Δcarbohydrate intake 8 ± 19 vs −18 ± 26 g/day, P = 0.43; Δprotein intake −4 ± 10 vs −8 ± 8 g/day, P = 0.75, pitavastatin vs placebo, respectively). Similarly, changes in activity level over the study were not different between pitavastatin and placebo (Δdaily metabolic equivalent tasks −29 ± 14 vs −9 ± 5, P = 0.17).

Safety and adverse events

Adverse events are shown in Table 3. There were two serious adverse events in the pitavastatin group and one in the placebo group. One participant in the treatment group was hospitalized overnight for acute chest pain with normal cardiac enzymes and was found to have no evidence of an acute myocardial infarction. Another patient in the treatment group received a diagnosis of metastatic cancer as described previously upon completion of the study. A participant in the placebo group was hospitalized for acute perforated appendicitis. There were no confirmed cases of rhabdomyolysis. There were similar rates of nonserious side effects in both groups (Table 3).

Table 3.

Adverse Events

	Pitavastatin (n = 25)	Placebo (n = 25)
Serious adverse events	2	1
Nonserious adverse events	16	25
Gastrointestinal symptoms	5	8
Diarrhea or loose stools	3	7
Constipation	1	0
Nausea	0	1
Heartburn	1	1
Excess gas	1	0
Muscle symptoms	4	6
Muscle aches or cramps	4	4
Dark urine	0	1
Rhabdomyolysis	0	0
Elevated CPK level	0	1
Other	6	8
Fatigue	2	5
Headache	0	2
Worsening depression	1	1
Flulike symptoms	0	1
Back pain	0	1
Vivid dreams	1	0

Sensitivity analyses

Excluding the patient who developed metastatic cancer, glucose parameters and liver fat remained unchanged between groups, and changes in ALT (1 ± 2 vs 3 ± 2 IU/L, pitavastatin vs placebo, P = 0.45) and AST (2 ± 2 vs 2 ± 2 IU/L, P = 0.85) were also highly similar between groups.

Multivariable linear regression analyses were performed to adjust for the change in weight over the study, which was modestly but nonsignificantly reduced with pitavastatin. Adjusting for changes in weight during the study, there remained no changes in low-dose M (P = 0.20), high-dose M (P = 0.72), HbA1c (P = 0.22), fasting glucose (P = 0.30), 2-hour glucose (P = 0.84), Ra glucose (P = 0.93), and liver fat content (P = 0.68) with pitavastatin compared with placebo, and HOMA-IR showed a similar small decrease with pitavastatin compared with placebo (P = 0.03).

In the subset of patients with ≥5% hepatic fat fraction at baseline (n = 32), changes in liver fat were not meaningfully different than in the entire cohort (Δhepatic fat fraction −1 ± 2 vs −2 ± 1%, pitavastatin vs placebo, P = 0.82).

Adherence to study medication

Adherence as assessed by pill counts at 12-week (97% ± 6% vs 95% ± 6%; pitavastatin vs placebo, P = 0.39) and final visits (96% ± 3% vs 96% ±2%, P = 0.94) was high.

Discussion

In this randomized clinical trial we demonstrated that pitavastatin treatment for 6 months at maximal clinical dosing of 4 mg daily did not affect measures of hepatic or whole-body insulin sensitivity, nor did it reduce liver fat, in men characterized by insulin resistance at baseline. We saw the expected change in LDL-C, and reported compliance was high, demonstrating good adherence to the study regimen.

The effects of statins on glucose appear to differ by intensity and individual formulation. Whereas higher-intensity statins such as atorvastatin and rosuvastatin when used at higher dosages show consistent effects on glucose levels (7, 9), lower-intensity statins such as pravastatin are not consistently shown to worsen glycemia (30–32). Meta-analyses of studies comparing moderate-intensity pitavastatin with either placebo or another statin have suggested no adverse effect on glucose metabolism (20), and a secondary prevention study comparing pitavastatin with atorvastatin showed beneficial effects of pitavastatin on glucose homeostasis (33). In contrast, a retrospective observational study suggested that pitavastatin increased diabetes risk more than other statins (34). Importantly, with increasing clinician awareness of the effect of certain statins to increase glucose, observational studies may increasingly be confounded by indication, as clinicians opt for statins that are thought to have less effect on glycemia in patients at higher risk for diabetes. The current study was designed to rigorously assess the effects of pitavastatin on hepatic and whole-body insulin sensitivity by using stable isotope and euglycemic hyperinsulinemic clamp methods. We elected to use a placebo comparator to assess the effects of pitavastatin compared with untreated natural history rather than a comparator statin with known effects on glucose. This study extends data from a limited number of smaller studies (≤20 participants) of hypercholesterolemic patients investigating one-step clamp data in shorter-term studies (≤2 months) without simultaneous liver fat data (32, 35).

Our results with regard to the effect of pitavastatin on glucose homeostasis are generally neutral, with no changes in hepatic or whole-body insulin sensitivity, HbA1c, or fasting or postprandial glucose and a modest beneficial effect on HOMA-IR with a trend toward reduction in fasting insulin. To assess the possibility that these findings of neutrality may be caused by type 2 error, it is important to examine the 95% CIs around the effect size (Table 2). These 95% CIs exclude clinically meaningful positive or negative effects of pitavastatin on HbA1c, high-dose M, and Ra glucose. For fasting glucose and for low-dose M, the 95% CI of the effect size suggests the possibility of a type 2 error in missing a modest beneficial effect of pitavastatin. Both of these measures are largely reflections of hepatic insulin sensitivity, as is HOMA-IR, and a study with a larger sample size may have shown a small beneficial effect of pitavastatin on these measures. Two-hour glucose has a fair amount of variance, as expected, such that the 95% CI of the effect size does not exclude a modest effect in either direction but does exclude a detrimental effect with clinical importance. Overall, we interpret our data to demonstrate that pitavastatin does not worsen glycemia over 6 months of treatment in an at-risk group of higher-weight, insulin-resistant men. This finding is of clinical relevance because statins are often prescribed for such individuals, and these are detailed data from a randomized controlled trial in humans conducted with state-of-the-art clamp techniques. Thus, this study adds important information to the field.

The underlying mechanism through which statins exert their effect on blood glucose levels is poorly understood (36). One possible mechanism whereby statins may worsen glycemia is reduction in CoQ10, with associated reductions in mitochondrial function, which then cause dysglycemia (and statin-induced myopathy) (37). We did show a reduction in serum CoQ10 levels in the pitavastatin group, but it was not associated with adverse glycemic effects. Importantly, we do not have tissue levels of CoQ10, which may be more relevant with regard to mitochondrial function and tissue-level glucose metabolism.

Nonalcoholic fatty liver disease (NAFLD) is a health risk to the growing population of obese patients, specifically those with visceral adiposity, and pharmacologic strategies are needed to augment efforts at lifestyle modification. Recent evidence has demonstrated that statins can be safely used in patients with prediabetes and nonalcoholic steatohepatitis (NASH) (38), and there is some evidence in the literature that statins may improve features of NAFLD (16). A small study in HIV-infected patients treated with atorvastatin vs placebo suggested that atorvastatin may reduce hepatosteatosis as assessed by liver/spleen ratio (39), but a larger study of HIV-infected patients randomly assigned to rosuvastatin or placebo showed no effect on liver fat scores calculated from serum metabolic and liver function tests (40). Post hoc analyses of large statin trials have suggested that statins reduce ALT in patients who have elevated transaminases at baseline, suggesting possible NAFLD (41, 42). An open-label study of atorvastatin in patients with NASH demonstrated improvement in steatosis and inflammation on biopsy, although fibrosis worsened in some patients (43), and a randomized controlled trial of simvastatin (N = 10) vs placebo (N = 6) for 1 year showed no improvement in aminotransferases or histologic features of steatosis, inflammation, or fibrosis (44). A 2013 Cochrane review of available evidence concluded that statins may improve aminotransferase levels and ultrasound findings but that trials with larger samples sizes are needed (16). A more recent expert panel recommended use of statins as potentially beneficial for NAFLD and NASH in those who have comorbid cardiovascular disease (CVD) (17). To our knowledge, there are no prospective, randomized, placebo-controlled trials to date investigating whether statins reduce liver fat content as assessed with MRS. Our data show no sign of a reduction in liver fat content, and the 95% CI for the effect size excludes a highly meaningful change. We also saw no change in AST or ALT. Notably, we did not require NAFLD for participation, although baseline levels of liver fat content in our cohort were high, on average well above the threshold for steatosis of 5%. Subset analysis of those who did meet the 5% threshold showed no changes in liver fat content as well. We could not perform liver biopsy, although quantification of liver fat content by ¹H MRS correlates very well with histology assessment of steatosis (45). We do not have information on histologic changes in inflammation or fibrosis, and we cannot rule out an effect of statins on one or more of these parameters.

Our study has limitations that should be noted. Not all of our participants had an absolute clinical indication for statins; rather, our eligibility criteria were focused on finding patients at high risk of worsening insulin resistance for whom statin treatment would be reasonable. We included only men for purposes of avoiding the potential cofounding of sex steroids on metabolic indices in a physiology study; thus, our study cannot be generalized to women. Larger future studies including both sexes are needed to investigate whether sex modifies the effect of statins on glycemic parameters. Additionally, we did not include patients >65 years old; statins are certainly prescribed in this older age group, and our results cannot be generalized to that clinical context. If the dysglycemic effects of statins develop over several years, our 6-month study duration may not have been sufficient to see adverse effects. A longer randomized controlled trial, the Long-Term Effects of High-Dose Pitavastatin on Diabetogenicity in Comparison With Atorvastatin in Patients With Metabolic Syndrome, is under way in Korea and will investigate the effect of pitavastatin vs atorvastatin on HbA1c over 2 years (46). Finally, there were some limitations to our clamp and stable isotope assessment. Measuring suppression of Ra glucose during the low-dose clamp would provide a pure measure of insulin’s ability to suppress gluconeogenesis in the liver, but the dosage of insulin used during the low-dose phase of the clamp was high enough that some subjects had complete suppression of gluconeogenesis and a small degree of glucose utilization, such that our data for percentage suppression of gluconeogenesis were not interpretable in some subjects. Additionally, a longer duration of clamp procedure at each insulin dose may have further increased glucose utilization. Nonetheless, data for fasting Ra and for M during low- and high-dose insulin clamp support the conclusion that pitavastatin did not worsen overall glucose homeostasis.

Our study also has important strengths, including enrollment of a cohort with baseline insulin resistance and use of detailed physiologic methods to assess glucose homeostasis. Data from this prospective randomized controlled trial using these techniques significantly extend findings from previous studies that pitavastatin is largely neutral to glucose homeostasis, perhaps with small beneficial effects on some measures. Although a larger sample size would be needed to definitively state that glycemia is unchanged by pitavastatin, our data strongly suggest that pitavastatin does not worsen glycemia. We used a placebo comparator to fully assess pitavastatin’s effect on insulin sensitivity, as opposed to measuring a marginal effect when compared with another statin with known effects on glucose. We did not show an effect of pitavastatin in reducing liver fat content or improving transaminases, although longer studies with liver biopsy may be needed to fully assess possible effects of statins on steatohepatitis. Overall, our findings suggest that pitavastatin may be a good pharmacological option for patients with insulin resistance or prediabetes, at moderate CVD risk in whom a statin might be considered for primary prevention. Additional studies of those with higher CVD risk and absolute indications for statin therapy may also be useful.

Glossary

Abbreviations:

ALT

alanine aminotransferase

ASCVD

atherosclerotic cardiovascular disease

AST

aspartate aminotransferase

BMI

body mass index

CoQ10

coenzyme Q10

CPK

creatine phosphokinase

CVD

cardiovascular disease

HDL-C

High-density lipoprotein cholesterol

HOMA-IR

homeostasis model assessment of insulin resistance

I

insulin concentration

LDL-C

low-density lipoprotein cholesterol

M

glucose disposal

MGH

Massachusetts General Hospital

MRS

magnetic resonance spectroscopy

NAFLD

nonalcoholic fatty liver disease

NASH

nonalcoholic steatohepatitis

OGTT

oral glucose tolerance test

SAT

subcutaneous adipose tissue

VAT

visceral adipose tissue